Abstract
The bidispersive equation, accounting for second-order temporal dispersion in a lossless medium, is examined under a Galilei-like transformation, as well as a conditional ordinary Galilei transformation. In the former case, the roles of space and time are reversed by comparison to the application of the same transformation to the quantum mechanical Schrödinger equation. Such an invariance can result in envelope speeds that can assume values above or below the group speed of the medium; also, they can be negative. Conditional Galilei transformation results in a decrease in dimensionality.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
It is well known that the Galilean transformation $z^{\prime} = z - vt,\;t^{\prime} = t$ arises from the “subluminal” Lorentz transformation
The Galilean transformation changes Eq. (2) as follows:
Obviously, the free particle Schrödinger equation is not form-invariant under this transformation. The first term on the left-hand side requires an extra phase to be added to the wavefunction in order to achieve form-invariance. This problem can be resolved using an extension of a technique introduced by Kennard [1] and Darwin [2] in the early days of quantum mechanics. Specifically, given a solution to Eq. (2), the expression
Choosing a finite-energy solution $\varphi ({x,y,z,t} )$ to Eq. (2), replacing in it $z$ by $z - vt$ and using the resulting expression in Eq. (4) will result in a finite-energy, localized, and linearly traveling wave solution to the 3D Schrödinger equation. The “envelope function” $\varphi ({x,y,z - v\,t,t} )$ travels with the speed $v$. However, the phase term in Eq. (4) travels at the speed $v/2$. The explicit dependence of $\varphi ({x,y,z - v\,t,t} )$ on time signifies that the wave function $\psi (x,y,z,t)$ will eventually disperse.
To derive an illustrative example of a finite-energy wave function $\psi (x,y,z,t)$, consider the following specific axisymmetric solution to Eq. (2) in cylindrical coordinates $({\rho ,\phi ,z})$:
Here, ${t_0}$ is an arbitrary positive parameter with units of time. Using this expression in conjunction with Eq. (4), one obtains the finite-energy traveling wave wavefunction
This specific solution was first derived by Darwin [1]. As the envelope moves along the z- direction with speed v, it spreads out (both along $\rho $ and z) and its amplitude decreases. However, for large values of the parameter ${t_0}$ (essentially for a long pulse), the solution remains localized up to a time t of $O({{t_0}} ).$
A variant of the Galilean transformation, specifically $x^{\prime} = x - az,\;z^{\prime} = z,$ has been useful in connection with the parabolic equation
Although both the Schrodinger and the paraxial equations are not form-invariant under a Galilean transformation, this is not the case for other types of equations. It has been found that form invariance under a Galilean transformation is feasible for certain nonlinear equations governing wave propagation in temporally dispersive media. Among them are the Korteweg-de Vries and the Benjamin-Bona-Mahony equations [9].
Consider, next, the “superluminal” Lorentz transformation
It is our goal in this article to study the implications of the dual Galilean transformation to the bidispersive equation describing paraxial electromagnetic wave propagation in a lossless temporally dispersive medium. A derivation of this equation, the application of the dual Galilean transformation for both normal and anomalous dispersion, and illustrative examples are given in Sec. 2. Two Ansätzen that facilitate solutions to the bidispersive equation for both normal and anomalous dispersion are discussed and illustrated with specific examples in Sec. 3. It is shown in Sec. 4 that the bidispersive equation can be conditionally invariant under an ordinary Galilean transformation. However, this entails a decrease in dimensionality. Concluding remarks are provided in Sec. 5.
2. Dual Galilean transformation of the bidispersive equation in a temporally dispersive medium
2.1 Bidispersive equation
Electromagnetic wave propagation in a linear, homogeneous, transparent, dispersive medium is governed by the scalar equation
In this expression, $u(\vec{r},t)$ is a real scalar field (a component of the vector-valued electric or magnetic fields) and $\beta _{op}^2( - i\partial /\partial t) = {({\omega /c} )^2}{n^2}( - i\partial /\partial t),n( - i\partial /\partial t)$ being the index of refraction, is a real pseudo-differential operator. A physical interpretation of the latter is provided in the frequency domain; specifically,
For a physically convenient central radian carrier frequency ${\omega _0}$, the real field $u(\vec{r},t)$ is expressed as follows:
Here, $\varphi (\vec{r},t)$ is a complex-valued envelope function and ${v_{ph}} = {\omega _0}/\beta ({\omega _0})$ denotes the phase speed in the medium computed at the central frequency ${\omega _0}$. A formal introduction of Eq. (11) into Eq. (9) yields the following exact equation governing the envelope function $\varphi (\vec{r},t)$[3–7]:
Here, $\nabla _ \bot ^2$ denotes the transverse (with respect to $z$) Laplacian operator. Usually, at this stage in the study of wave propagation through dispersive media, one introduces the moving reference frame $\xi = z,\;\;\tau = t - (z/{v_{gr}})$, in terms of the group speed ${v_{gr}} = 1/{\beta _1};\;\;{\beta _1} \equiv { {d\beta (\omega )/d\omega } |_{\omega = {\omega _0}}}$. Then, Eq. (12) is transformed into the expression
Several techniques have been developed based on the type of approximations made to the exact Eq. (13). In the sequel, use will be made of the slowly varying envelope approximation (SVEA) (cf. e.g., [4]), whereby one neglects the second derivative with respect to z (paraxial approximation), as well as the mixed derivative term involving $z$ and $\tau $. Furthermore, dispersive effects up to the second order will be retained. One, then, has the bidispersive equation
2.2 Dual Galilean transformation of the bidispersive equation
Motivated by the dual Galilean transformation $t^{\prime} = t - z/v,\;z^{\prime} = z\;,$ the change of variables $\tau ^{\prime} = \tau - z/v,\;z^{\prime} = z\;\;$ will be applied to the bidispersive Eq. (14) without a restriction on the speed v. The bidispersive equation is not form-invariant under this transformation. A phase term must be added to the wavefunction for this requirement to be fulfilled. Specifically, given a solution to Eq. (14), the expression
Recall that in Eq. (14) $\tau = t - z/{v_{gr}}.$ It is clear, then, as mentioned above, that the envelope wavefunction no longer is seen to travel at the medium group velocity ${v_g}.$ Instead, one obtains $\tau - \frac{z}{v} = T = t - \frac{z}{{{v_{en}}}},$ with ${v_{en}} = v\,{v_g}{({v + {v_g}} )^{ - 1}}$ the new envelope speed. For $v > 0,$ one has ${v_{en}} < {v_g}.$ In the limiting case $v \to \infty ,$ one obtains ${v_{en}} \to {v_g}.$ Finally, ${v_{en}} > {v_g}$ for $v < - {v_g},$ but the wavepacket moves backwards.
In addition to the new envelope speed, a new phase speed is introduced. From Eq. (15), one obtains $\left( {\tau - \frac{z}{{2v}}} \right) = t - \frac{z}{{{V_{ph}}}},$ where ${V_{ph}} = 2v\,{v_g}{({{v_g} + 2v} )^{ - 1}}$ is the new phase speed. As in the case of the envelope speed, ${V_{ph}}$ is smaller from ${v_g}$ for $v > 0$. A combination of the phase factors in Eqs. (11) and (15), gives rise to an overall effective phase speed $V_{ph}^{eff} = {v_{ph}}\left( {{\beta_0} + \frac{1}{{{{\bar{\beta }}_2}v{v_{ph}}}}} \right){\left( {{\beta_0} + \frac{1}{{{{\bar{\beta }}_2}v{V_{ph}}}}} \right)^{ - 1}}.$
2.3 Illustrative example: normal dispersion
Consider the (3 + 1) D bidispersive equation
Here, ${a_1}\;\textrm{and }a$ are positive parameters, the latter small enough to guarantee the square integrability of the solution. Next, the dual Galilean transformation in Eq. (17) is carried out. Specifically,
Plots of $|{\varphi ({0,z,\tau } )} |$ and $|{\psi ({0,z,\tau } )} |vs.\textrm{ }\tau \textrm{ and }z$ are shown in Figure 1.
2.4 Illustrative example: anomalous dispersion
A specific solution of Eq. (14) in the case of anomalous dispersion is given by
3. Two ansätzen for the bidispersive equation
It will be convenient for the following work to introduce the dimensionless variables
In this manner, the diffraction length ${L_{diff}} = {\beta _0}x_o^2$ is equal to the corresponding dispersion length ${L_{disp}} = {\tau ^2}/{\bar{\beta }_2}.$ Here, ${x_0}$ is an arbitrary length scale and ${\tau _0} = {x_0}\sqrt {{\beta _0}{{\bar{\beta }}_2}} $ is related to the pulse width of the wavepacket. We have, then, in the place of Eq. (14)
3.1 Ansatz I. Normal dispersion
Consider any solution of the Klein-Gordon equation
Then,
With $q = 1 + {a^2}/2 + b,$ we obtain the ansatz given in Eq. (24).
3.1.1 Illustrative example
A nonsingular azimuthally symmetric solution to the Klein-Gordon equation in (23) is given by
This is an infinite-energy envelope-invariant solution propagating at the dimensionless speed $Z/T = V = 1/a.$ A finite-energy solution can be derived by making use of the free parameter $b.$ One such solution in terms of the original variables and with $a = ({{x_0}/v} ){({{\beta_0}/{{\bar{\beta }}_2}} )^{1/2}}$ is given as follows:
Here, $erfc({\cdot} )$ denotes the complementary error function. The envelope and phase speeds are those specified earlier. Fig. 3 shows surface plots of $|{\psi (\rho ,\varsigma } |\textrm{ vs}\textrm{. }\varsigma = z - {v_{en}}\textrm{t and }\rho $ and ${\textrm{Re}} \{{\psi ({\rho ,\varsigma } )} \}\textrm{ vs}\textrm{. }\varsigma $ for $\rho = 0$.
3.2 Ansatz II. Anomalous dispersion
Consider a solution of the Helmholtz equation
Then,
3.2.1 Illustrative example
A nonsingular azimuthally symmetric solution to the Helmholtz in (30) is given by
Application of the ansatz in Eq. (31) yields the solution
This is an infinite-energy envelope invariant solution propagating at the dimensionless speed $V = 1/a.$ A finite-energy solution can be derived by making use of the free parameter $b.$ One such solution in terms of the original variables and with $a = ({{x_0}/v} ){({{\beta_0}/{{\bar{\beta }}_2}} )^{1/2}}$ is given as follows:
The envelope and phase speeds are those specified earlier. Fig. 4 shows $|{{\textrm{Re}} \{{\psi ({\rho ,\varsigma } )} \}} |\textrm{ vs}\textrm{. }\varsigma = z - {v_{en}}\textrm{t and }\rho $ and ${\textrm{Re}} \{{\psi ({\rho ,\varsigma } )} \}\textrm{ vs}\textrm{. }\varsigma $ for $\rho = 0$.
4. Conditional Galilei invariance of the bidispersive equation
An interesting question is whether the bidispersive equation can be invariant under an ordinary Galilean transformation. Let $T = \tau $ and $Z = z - v\tau .$ With these changes of coordinates, the bidispersive equation (14) changes to
It follows, then, that for ordinary Galilei invariance, the right-hand side of Eq. (32) must vanish. It turns out that this is the case for any function $\varphi ({\rho ,\phi ,Z + vT/2} ).$ We consider, then, the left-hand side of Eq. (35) which can be written as
4.1 Illustrative example: normal dispersion
We present, next, a specific transformed solution to the bidispersive equation:
This represents an infinite-energy X wave [14–18]. Both the envelope and phase speed are equal to $v\,{v_g}/({v + 2{v_g}} )$.
4.2 Illustrative example: anomalous dispersion
In the case of anomalous dispersion, Eq. (36) becomes
A specific solution to this equation is given by
The infinite-energy X wave given in Eq. (37) and the infinite-energy MacKinnon -like structure in Eq. (39) are depicted below in Figs. 5(a) and 5(b), respectively.
5. Concluding remarks
The bidispersive equation, accounting for second-order temporal dispersive effects in a homogeneous lossless medium, has been examined in detail, first with respect to a Galilei transformation dual to that used in connection to the quantum mechanical Schrödinger equation. Furthermore, two Ansätzen have been introduced that allow the derivation of large classes of solutions of the bidispersive equation. Specific illustrative examples are given for both normal and anomalous dispersion. Finally, an investigation of a conditional ordinary Galilei transformation has revealed that it results in a decrease in dimensionality.
A note on superluminality underlying the dual Galilean transformation is appropriate. The scalar wave equation and, more generally, Maxwell’s equations in free space are invariant under the superluminal Lorentz transformations given in Eq. (8). The presence of a superluminal speed in finite-energy solutions to these equations does not contradict relativity. If parameters entering the solutions are chosen appropriately, a pulse moves superluminally with almost no distortion up to a certain distance, say ${z_d} = vt,$ and then it slows down to a luminal speed c, with significant accompanying distortion. Although the peak of the pulse does move superluminally up to ${z_d},$ it is not causally related at two distinct ranges ${z_1},{z_2} \in [{0,{z_d}} )$. Thus, no information can be transferred superluminally from ${z_1}$ to ${z_2}$.
The discussion in this article has been confined mostly to the implications of the dual Galilean transformations to a bidispersive equation. However, as in the case of the quantum mechanical Schrödinger equation whose solutions are invariant under other types of symmetry transformations, e.g., a conformal transformation, a dual such transformation applies to the bidispersive equation. Specifically, given a solution ${\varphi _ \mp }({x,y,\iota ,z} )$ of the bidispersive Eq. (14) for normal and anomalous dispersion, respectively, the conformal transformation
Although the bidispersive equation is not form invariant under a dual Galilei transformation, a specific solution to the bidispersive equation remains a solution under such transformation. The same applies, for example, to the energy conservation law. Given the dimensionless bidispersive Eq. (22), the energy conservation laws for anomalous and normal dispersion are given as flows:
They are applicable for both the solutions of the bidispersive equation and the transformed ones.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. E. H. Kennard, “Zur Quantenmechanik einfacher Bewegunstypen,” Z. Phys. 44(4-5), 326–352 (1927). [CrossRef]
2. C. G. Darwin, “Free motion in quantum mechanics,” Proc. R. Soc. Lond. A 117(776), 258–293 (1927). [CrossRef]
3. T. Brabec and F. Krausz, “Nonlinear optical pulse propagation in a single-cycle regime,” Phys. Rev. Lett. 78(17), 3282–3285 (1997). [CrossRef]
4. M. A. Porras, “Ultrashort pulsed Gaussian light beams,” Phys. Rev. E 58(1), 1086–1093 (1998). [CrossRef]
5. M. A. Porras, “Propagation of single-pulse light beams in dispersive media,” Phys. Rev. A 60(6), 5069–5073 (1999). [CrossRef]
6. P. Kinsler and G. H. C. New, “Few-cycle pulse propagation,” Phys. Rev. A 67(2), 023813 (2003). [CrossRef]
7. X. Fu, L. Qian, S. Wen, et al., “Propagation of the ultrashort pulsed beam with ultrabroad bandwidth in the dispersive medium,” Phys. Rev. A 68(6), 063818 (2003). [CrossRef]
8. G. A. Siviloglou and D. N. Christodoulides, “Accelerating finite energy Airy beams,” Opt. Lett. 32(8), 979–981 (2007). [CrossRef]
9. A. Duran, D. Dutykh, and D. Mitsotakis, “On the Galilean invariance of some dispersive wave equations,” Stud. Appl. Math. 131(4), 359–388 (2013). [CrossRef]
10. J. M. Lévy-Leblond, “Une nouvelle limite non-relativiste du groupe de Poincaré,” Ann. Inst. Henri Poincaré, Sect. A 3(1), 1–12 (1965).
11. N. S. Sen Gupta, “On an analogue of the Galilei Group,” Il Nuovo Cim. XLIV, 512–517 (1966).
12. A. Chong W, H. Renninger, D. N. Christodoulides, et al., “Airy-Bessel wave packets as versatile linear light bullets,” Nat. Photonics 4(2), 103–106 (2010). [CrossRef]
13. N. K. Efremidis, Z. Chen, M. Segev, et al., “Airy beams and accelerating waves: An overview of recent advances,” Optica 6,686 (2019). [CrossRef]
14. J. Y. Lu and J. F. Greenleaf, “Nondiffracting X Waves - Exact Solutions to Free Space Scalar Wave Equation and their Finite Aperture Realization,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr. 39(1), 19–31 (1992). [CrossRef]
15. P. Saari and K. Reivelt, “Evidence of X-Shaped Propagation-Invariant Localized Light Waves,” Phys. Rev. Lett. 79(21), 4135–4138 (1997). [CrossRef]
16. I. M. Besieris, M. Abdel-Rahman, A. M. Shaarawi, et al., “Two Fundamental representations of localized wave pulse solutions to the scalar wave equation,” Progress in Electromagn. Res. PIER 19, 1–48 (1998). [CrossRef]
17. J. Salo, J. Fagerholm, A. T. Freiberg, et al., “Unified description of nondiffracting X and Y waves,” Phys. Rev. E 62(3), 4261–4275 (2000). [CrossRef]
18. P. Saari and K. Reivelt, “Generation and classification of localized waves by Lorentz transformations in Fourier space,” Phys. Rev. E 69(3), 036612 (2004). [CrossRef]
19. H. E. Hernandez-Figueroa, M. Zamboni-Rached, and E. Recami, Localized Waves (Wiley-Interscience, 2008).
20. H. E. Hernandez-Figueroa, E. Recami, and M. Zamboni-Rached, Non-Diffracting Waves (Wiley-VCH, 2014).